Various types of apparatus have been developed or proposed for recording a pattern of pixels onto a photosensitive medium, using various types of light sources including LEDs and lasers. In the conventional model, image content is recorded onto photosensitive media, such as photosensitive film or paper, a full frame at a time. A number of pixel-based digital imaging apparatus follow this traditional model by modulating a full frame of pixels at one time for exposure of the image content. For example, various types of two-dimensional spatial light modulators, such as liquid crystal devices (LCDs) or digital micromirror devices (DMDs) can be used to provide a complete frame of image data for exposure.
As just one example, commonly-assigned U.S. Pat. No. 6,215,547 (Ramanujan et al.) discloses a writing apparatus employing a reflective LCD spatial light modulator for providing modulated light for exposure of a photosensitive medium, one image frame at a time.
In applying this conventional frame-based imaging model, the photosensitive medium is moved or indexed into position within an exposure apparatus and is then maintained in a stationary position during exposure of the pixel pattern within the image frame. The job of exposing successive pixel image frames onto a length of photosensitive medium requires successive steps for moving and stopping the media to record each frame.
It can be appreciated that constant starting and stopping of media movement has a number of drawbacks, particularly with respect to throughput and to the overall mechanical complexity of the film transport system. In response to the need for improved efficiency, a number of alternatives have been proposed. For example:                U.S. Pat. No. 6,163,363 (Nelson et al.) discloses a DMD spatial light modulator used to expose an image onto a continuously moving photosensitive medium, one or more lines of pixels at a time.        Similarly, U.S. Pat. No. 5,953,103 (Nakamura) discloses a color printer using an array of modulated light sources that records four lines of pixels at a time by progressively indexing the media past a stationary printhead.        U.S. Pat. No. 5,968,719 (Nakamura) discloses a side printer for printing bar codes and other information onto a section of filmstrip media during processing.        
While the above-listed patents describe methods for writing one or more lines of pixels onto moving photosensitive media, these methods are limited to applications in which the photosensitive medium moves through the exposure region at a relatively constant speed. There is a need to print digital watermark images onto motion picture photosensitive medium while the photosensitive medium is in motion. Such would be the situation in the manufacturing process of the motion picture photosensitive medium where forming latent watermarks images on the photosensitive medium would be done while the photosensitive medium was moving at high speeds. The prior art methods listed above would not be readily suitable for applications in which the photosensitive medium moves at variable speeds.
Addressing the problem of writing pixels at variable media speed, commonly-assigned U.S. Pat. No. 5,294,942 (Loewenthal et al.) discloses an apparatus for forming a pixel pattern, one line of pixels at a time, onto a medium that is moving at a variable rate. The apparatus of U.S. Pat. No. 5,294,942 tracks the speed of the moving photosensitive medium and adapts its pixel exposure timing, based on speed tracking results, to obtain a uniform exposure. The method and apparatus of U.S. Pat. No. 5,294,942 thus provides a more flexible solution for obtaining uniform exposure levels for recording pixels. For example, a pattern of pixels can be recorded on the leading or trailing end of a film roll without requiring that the film be moving through an exposure region at a constant speed.
However, while methods described in U.S. Pat. No. 5,294,942 and in related prior art enable the recording of a pixel pattern onto a continuously moving medium in line-by-line fashion, high-speed manufacturing and film processing environments can impose even further requirements. One area of particular concern relates to forming a latent image watermark onto a photosensitive medium during manufacture of the medium.
For example, as is disclosed in U.S. Patent Application 2003/0012569 (Lowe et al.), a latent watermark image can be exposed onto the “raw” photosensitive medium itself, at the time of manufacture. Then, when the medium is exposed with image content, the image frame is effectively overlaid onto the watermark pattern. Such a method is also disclosed in U.S. Pat. No. 6,438,231 (Rhoads). The Rhoads '231 patent discloses this type of pre-exposure of the watermark onto the film emulsion within the frame area of negative film, for example.
It can be appreciated that watermark pre-exposure would have advantages for marking motion picture film at the time of manufacture or prior to exposure with image content. A length of motion picture film could be pre-exposed with unique identifying information, encoded in latent fashion, that could be used for forensic tracking of an illegal copy made from this same length of film. However, prior art watermarking techniques proposed for photosensitive media in general fall short of what is needed for motion picture watermarking, particularly watermarking during high-speed film manufacture. Problems that make it difficult or impractical to use conventional watermark application techniques for pre-exposure of film in manufacture relate to both throughput requirements and image quality. Among the problems with watermark application in high-speed manufacturing environments are the difficulty of exposure control, not only for maintaining a uniform exposure, but for modulating exposure to produce a watermark pixel pattern having a selectable number of grayscale levels. Another problem, not a factor during pixel-wise exposure at lower speeds, relates to pixel shape. That is, with the photosensitive medium moving at high speeds during pixel recording, there can be a significant amount of pixel elongation in the travel direction, visible as “smear.” Unfortunately, the amount of pixel smear varies with the speed of media travel, effectively changing the dimensions of the pixel depending on the specific rate of speed of the media past the exposure source.
Referring to FIG. 1A, there is represented how a pixel 10 is recorded onto a photosensitive medium 12 by a pixel exposure source 14 when photosensitive medium 12 is stationary. (FIG. 1A elements are not to scale, but are represented to show the overall concept.) The exposure light beam from pixel exposure source 14 has a uniform power output density W (typically expressed in Watts/cm2). This exposure level is enabled for a period of time (t), or exposure time, to create a density (D) where D=log H. As is well known in the imaging arts, exposure (H) is a function expressed in general terms as H=W−t. The overall shape of pixel 10 resembles the output shape of pixel exposure source 14; a circular output aperture of pixel exposure source 14 yields a substantially circular pixel 10. In FIG. 1A, pixel exposure source 14 has an output diameter of some arbitrary pixel size, depending on the application. Since pixel dimensions can vary over a range, the pixel diameter is simply considered as a normalized “pixel unit” in the description that follows. With reference to FIG. 1A, latent image pixel 10 formed by exposure is 1 pixel unit in diameter and has a density level, D. As shown in FIG. 1B, a density profile 32 in the direction along the length of the medium through the center of the pixel has a uniform density profile 32, so that density D is fairly consistent across pixel 10.
In contrast with FIG. 1A, FIG. 2A shows how an elongated pixel 20 is formed when photosensitive medium 12 is moving, in the direction of the arrow. Again, FIG. 2A is not to scale, but is sized for comparison with FIG. 1A. For example, photosensitive medium 12 is transported in a length direction, termed its travel direction, at a velocity of V mm/t, during exposure time t. The resulting exposure on photosensitive medium 12 forms an elliptical pixel 20 with a dimension that is a factor of 1 pixel unit times V velocity. The elongated or elliptical shape of pixel 20, also termed “smearing,” is caused by the movement of photosensitive medium 12 while pixel 20 is exposed. A density profile 32a through the center of pixel 20 shows non-uniform density, as is shown in FIG. 2B. This non-uniformity of density occurs since the middle third of the mid section receives light for the full exposure time t while leading and trailing portions of pixel 20 receive light for a shorter time, which can be considered as the integrated time t/2.
FIGS. 3A and 3B show a representative portion of a pixel pattern where photosensitive medium 12 is held stationary and where moving, respectively. Different shading is used to indicate that each individual pixel 10, 20 is also assigned a density level D. By comparing FIGS. 3A and 3B, it is readily seen that different pixel 10, 20 shapes result, depending on whether or not photosensitive medium 12 is moving. Pixels 20 also exhibit a different density profile 32a depending on speed. Moreover, both the shape of pixel 20 in the travel direction and its density profile 32a will vary depending on the transport speed of photosensitive medium 12. Thus, the change in pixel 20 shape and density profile 32a with transport speed complicates the task of forming latent indicia of any type onto photosensitive medium 12. In addition, variation in pixel 20 shape and density profile 32a also make it difficult to modulate the relative density of pixel 20 to allow encoding of information corresponding to pixel 20 density.
Where the speed of photosensitive medium 12 is fairly slow, the actual effect of pixel 20 smearing, as represented in FIG. 2A, is negligible. That is, the exposure time is so short that the basic response represented in FIG. 1A occurs for slow moving photosensitive medium 12. On the other hand, the faster the speed of photosensitive medium 12, the more pronounced is the elongation of pixel 20. It can be appreciated by those skilled in the imaging arts that forming pixels 20 at very high film speeds can result in considerable distortion of pixel 20. It can be difficult to control both the shape and the effective density of pixel 20, particularly if the rate of photosensitive medium 12 speed changes. For instance, at the beginning or near the end of a spool of a film medium, the necessary acceleration or deceleration of the film medium would cause pixels 20 to have different dimensions relative to their dimensions at portions of the film medium when printed at full speed. Moreover, any attempt to control the density level of pixels 20 recorded at various media transport speeds would be particularly difficult using existing exposure timing techniques. Thus, it would be very difficult to record, in high-speed film manufacture or processing, a pattern of pixels 20 having consistent shape and having controllable effective density over all portions of a length of photosensitive medium 12.
FIGS. 3B and 3C show the elongated nature of pixels 20 and show how some amount of overlap can occur between adjacent pixels 20 in the length direction. FIG. 3B shows the spatial outlines of pixels 20 in dotted line form, with only two pixels 20 represented to show non-uniform density profiles 32a, corresponding to two pixels 10 in FIG. 3A. The overlap area between pixels 20 may effectively receive exposure for two pixels 20; however, the effect on density in this overlap area has been shown to be minimal, due to response characteristics of the media. FIG. 3C shows the inter-pixel timing in more detail, with a preferred timing of encoder pulses 28, as described subsequently, and resulting density profiles 32a for each successive pixel 20 in the length direction. The combined effect of the applied exposure is shown as a pixel-to-pixel additive density profile 44. Again, the additive density profile 44 shown in FIG. 3C is a first approximation; response characteristics of the media are a key factor in determining the effects of successive applications of exposure energy from different exposure sources in the overlap area of pixels 20.
From an imaging perspective, properties of a watermark image or other indicia may need to meet high levels of quality. For example, complex watermark extraction methods may require that certain properties of watermark pixels be maintained in order to allow successful extraction of the encoded information. Pixel size and density are among key properties for this purpose. Thus, there is a need for methods of forming pixel patterns for watermarks and other latent indicia on photosensitive media, where the method compensates for acceleration/deceleration of the moving medium.